RE: [Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers
Thankyou all. Once I get more familiar with Parsec I might do my own lexeme parser. For now I've changed its definition of simpleSpace to simpleSpace = skipMany1 (satisfy (\c - (c /= '\n') (isSpace c))) so that it skips whitespace, but not newlines. My parser then explicitly matches newlines whenever I need to. Cheers, Paul -Original Message- From: Brandon S. Allbery KF8NH [mailto:[EMAIL PROTECTED] Sent: 26 March 2008 11:48 To: haskell-cafe@haskell.org Cafe; Paul Keir Subject: Re: [Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers On Mar 26, 2008, at 7:42 , Bulat Ziganshin wrote: Wednesday, March 26, 2008, 2:32:53 PM, you wrote: I'm looking to parse a Fortran dialect using Parsec, and was afair, some months ago BASIC parsing was discussed here. the first solution one can imagine is to add preprocessing stage replacing line ends with ';'-alike FWIW I just ignored the lexeme parser and did my own on top of the basic Parsec primitives. You may need to do that anyway if you want to support older variants of Fortran, which don't actually have keywords and ignore spaces outside of string constants (Hollerith constants) --- Parsec's lexeme stuff doesn't even pretend to support this. -- brandon s. allbery [solaris,freebsd,perl,pugs,haskell] [EMAIL PROTECTED] system administrator [openafs,heimdal,too many hats] [EMAIL PROTECTED] electrical and computer engineering, carnegie mellon universityKF8NH ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers
Hi, I'm looking to parse a Fortran dialect using Parsec, and was hoping to use the ParsecToken module. Fortran though is unlike the Parsec examples, and prohibits default line continuation (requiring instead an explicit ampersand token ). The lexical parsers of the ParsecToken module skip whitespace after each symbol parsed (lexeme parsers); including the newline. I was thinking of making a minor change to Parsec: In Token.hs, lexeme, whiteSpace, and simpleSpace are defined. lexeme uses whiteSpace which uses simpleSpace. simpleSpace is defined: simpleSpace = skipMany1 (satisfy isSpace) I could replace this locally with: simpleSpace = skipMany1 (satisfy isSpacePlusAmpersandMinusNewline) This approach may have other problems though. Has anyone experience of using Parsec to parse such a language? For example assembly or a preprocessor? Thanks in advance, Paul ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers
Hello Paul, Wednesday, March 26, 2008, 2:32:53 PM, you wrote: I'm looking to parse a Fortran dialect using Parsec, and was afair, some months ago BASIC parsing was discussed here. the first solution one can imagine is to add preprocessing stage replacing line ends with ';'-alike -- Best regards, Bulatmailto:[EMAIL PROTECTED] ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers
Paul Keir wrote: Hi, I'm looking to parse a Fortran dialect using Parsec, and was hoping to use the ParsecToken module. Fortran though is unlike the Parsec examples, and prohibits default line continuation (requiring instead an explicit ampersand token ). The lexical parsers of the ParsecToken module skip whitespace after each symbol parsed (lexeme parsers); including the newline. If you mean Text.ParserCombinators.Parsec.Token, can't you just change the setting of whiteSpace in TokenParser? ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] No newlines in the whitespace for Parsec lexeme parsers
On Mar 26, 2008, at 7:42 , Bulat Ziganshin wrote: Wednesday, March 26, 2008, 2:32:53 PM, you wrote: I'm looking to parse a Fortran dialect using Parsec, and was afair, some months ago BASIC parsing was discussed here. the first solution one can imagine is to add preprocessing stage replacing line ends with ';'-alike FWIW I just ignored the lexeme parser and did my own on top of the basic Parsec primitives. You may need to do that anyway if you want to support older variants of Fortran, which don't actually have keywords and ignore spaces outside of string constants (Hollerith constants) --- Parsec's lexeme stuff doesn't even pretend to support this. -- brandon s. allbery [solaris,freebsd,perl,pugs,haskell] [EMAIL PROTECTED] system administrator [openafs,heimdal,too many hats] [EMAIL PROTECTED] electrical and computer engineering, carnegie mellon universityKF8NH ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe